NoSQL

 A database management system designed to handle and store large volumes of

unstructured and semi-structured data.
 It uses flexible data models unlike traditional relational databases and supports
horizontal scaling.
 Originally referred to "non-SQL" or "non-relational," now meaning "not only SQL,"
reflecting a broader range of database architectures.
When to Use NoSQL
 When managing large volumes of data that grows continuously.
 When relationships between data are not a key concern.
 When data is unstructured or semi-structured and evolves over time.
 When database-level support for constraints and joins is unnecessary.
 When high scalability and performance for large-scale data retrieval and storage are
required.
Advantages
 High availability
 High scalability
 flexibility
Disadvantages
 Lack of ACID compliance
 Lack of standardization
 Large document size
Cap theorem

The CAP Theorem says that in a distributed system, you can only guarantee two out of three
things:
 1. Consistency (C): Every node sees the same data at the same time.
2. Availability (A): Every request gets a response, even if some parts of the system are
down.
3. Partition Tolerance (P): The system keeps working even if parts of it can’t
communicate with each other.
 CA: Guarantees consistency and availability, but no tolerance for network issues.
 CP: Guarantees consistency and partition tolerance, but may sacrifice availability.
 AP: Guarantees availability and partition tolerance, but may have inconsistent data
during network issues.
ACID Transactions (Simplified)
ACID is a set of rules that ensure data is handled correctly in a database, especially when
making changes.
1. Atomicity:
o Everything in a transaction happens together or nothing happens. If
something goes wrong, all changes are undone.
o Example: If you transfer money between accounts, it either happens fully or
not at all.
2. Consistency:
o The database must always be in a valid state. After a transaction, all rules (like
no negative balances) must be followed.
o Example: If a rule says no account can go below zero, the transaction won’t
complete if it causes that.
3. Isolation:
o Transactions don’t interfere with each other. Changes made in one
transaction are not seen by others until it's fully done.
o Example: If two people are transferring money at the same time, one’s
change won’t overwrite the other.
4. Durability:
o Once a transaction is completed, its changes are saved permanently, even if
the system crashes.
o Example: After you transfer money, the record stays in the system even if the
computer crashes.
ACID ensures data stays correct, safe, and consistent during transactions.
Aggregate Data models
 In NoSQL databases, aggregate data models refer to the way data is grouped or
organized into collections or units that are treated as a single entity for storage and
retrieval.
 This model is designed to improve performance and make it easier to handle
complex data structures.
Types of Data Models in NoSQL Databases
1. Column-Based Database

 Description:
o Stores data in tables with rows and flexible columns.
o Different rows can have different columns, making it highly adaptable.
o Uses compression for space-saving and improved speed.
o Suitable for retrieving large amounts of sparse data.
 Example:
o Apache HBase: A distributed database designed to handle large-scale data
quickly.
 Advantages:
o Highly efficient data compression and partitioning.
o Scalable and suitable for big data and data warehousing.
 Disadvantages:
o Inefficient for updates and deletions due to column-oriented storage.
o Slower performance for queries involving multiple attributes.
2. Document-Oriented Databases
 Description:
o Stores data in JSON-like documents containing field-value pairs.
o Supports complex and nested data structures, ideal for unstructured data.
 Example:
o MongoDB: A document-based database supporting replication and dynamic
schemas.
 Advantages:
o Flexible schema allows easy modification of data structures.
o Fast retrieval and storage of hierarchical data.
o Good for applications requiring varying or unstructured attributes.
 Disadvantages:
o No support for complex transactions across multiple documents.
o Relationships between data are not well-defined.
[Link]-Based Database
 Description:
o Stores data as nodes (entities like people or places) and edges (relationships).
o Designed to uncover patterns in highly interconnected data.
 Example:
o Neo4j: A popular graph database for analyzing complex relationships.
 Advantages:
o Excellent for traversing and querying relationships between data.
o Flexible schema for dynamic data models.
o Suitable for applications like recommendation engines and fraud detection.
 Disadvantages:
o Difficult to scale for large datasets.
o Lack of a standard query language.
[Link]-Value Databases
  Description:
o Simplest type of NoSQL database, storing data as key-value pairs.
o Optimized for fast data retrieval, commonly used for caching and session
management.
 Example:
o Amazon DynamoDB: A widely-used key-value database that popularized
NoSQL.
 Advantages:
o Extremely fast due to its simple structure.
o Easy to scale horizontally across distributed systems.
o No complex query optimization required.
 Disadvantages:
o Limited to basic operations like GET, PUT, and DELETE.
o No support for complex queries, joins, or foreign key constraints.
5. MongoDB
 Definition: Open-source document database with dynamic schema, high
performance, and scalability.
 Key Features:
o Stores data in BSON documents.
o Supports ad-hoc queries, replication, and sharding.
o Schema-less and JSON-like storage.
 Sharding Components:

o Shards (replicas): Store data and ensure consistency.

 o Configuration Servers: Manage cluster metadata.
o Query Routers: Interface for applications to query data.
 Advantages:
o Flexible and schema-less.
o Easy to scale.
o Supports indexes and replication.
o Faster due to internal memory usage.
o Can function as a file storage system (GridFS).
 Disadvantages:
o Limited for complex transactional systems.
o No explicit database creation commands.
6. Schemaless Databases in NoSQL
Schemaless databases in NoSQL don't require a fixed structure or schema for storing data.
They allow data to be stored and retrieved without following a specific format, offering
flexibility for changing data over time.
How They Work
 Storage Types:
o On-the-disk: Uses B-Trees with the top stored in RAM.
o In-memory: Uses RB-Trees, with only data appends stored on disk.
 No Fixed Schema:
o No strict rules for data structure. The database adapts as needed.
Data Storage
 Data is stored in JSON-style documents where each document can have different
fields and data types.
o Example:
o {name: "Joe", age: 30, interests: 'football'}
o {name: "Kate", age: 25}
 Even though it's "schemaless," databases like MongoDB still have some structure
(e.g., collections and indexes).
Types of NoSQL Databases Supporting Schemaless Models
 Key-Value Store
  Document Store (e.g., MongoDB)
 Column-Oriented Database
 Graph Databases
Benefits
1. Flexibility
2. Scalability
3. Simplicity
Disadvantages
1. Inconsistencies
2. Management Complexity